junte/Model
1
0
Fork 0
mirror of https://github.com/Hestia-Homes/Model.git synced 2026-06-08 11:17:27 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Slice S0380.176: Table 4b combi sub-row dispatch for (61)m

Browse source 
SAP 10.2 §4 line 7702 (PDF p.137) defines (61)m as "Combi loss for
each month from Table 3a, 3b or 3c (enter '0' if not a combi
boiler)". Table 4b sub-rows 128 / 129 / 130 are explicit combi sub-
rows per the spec row names:
    128: Combi oil boiler, pre-1998
    129: Combi oil boiler, 1998 or later
    130: Condensing combi oil boiler

Pre-slice `_table_3a_combi_loss_default_applies` gated only on
`main_heating_category ∈ {1, 2, 3, 6}`. The Elmhurst mapper leaves
`main_heating_category=None` on Table 4b liquid-fuel boilers (FAME,
HVO, B30K) — the cascade fell through to (61)m=0 despite the lodged
SAP code being a combi sub-row, under-counting (62)m by 600 kWh/yr
for FAME combi certs.

Extended the helper with a `_TABLE_4B_COMBI_OR_CPSU_CODES` fall-
through (set already exists for the symmetric `_primary_loss_
applies` Table 4b non-combi branch — see S0380.146). The set carries
the canonical combi + CPSU sub-row codes (103/104/107/108/112/113/
118/120-123/128-130). For cylinder-lodged certs the existing
`if epc.has_hot_water_cylinder: combi_loss_override = zero_monthly`
guard in `_water_heating_worksheet_and_gains` still pre-empts the
combi-loss fall-through correctly — non-combi codes with cylinders
remain (61)m=0.

Closures (heating-systems corpus 001431):
  oil 3 (code 128, FAME, no cylinder) ALL EXACT (±0.0000):
    ΔSAP_c +2.5863 → -0.0000
    Δcost  -£61.89 → -£0.00
    ΔCO2   -14.58  → +0.00
    ΔPE    -967.10 → +0.00
  oil 4 (code 129, FAME, no cylinder) ALL EXACT (±0.0000):
    ΔSAP_c +2.5603 → +0.0000
    Δcost  -£56.66 → +£0.00
    ΔCO2   -13.35  → +0.00
    ΔPE    -884.90 → +0.00

Oil 6 (code 126, NOT a combi, with cylinder) unchanged — the fix
is gated on the combi sub-row set. Cohort moves from 9 pinned
residuals to 7.

933 pass + 0 fail (+1 new mapper test). Pyright net-zero on cert_
to_inputs.py + tests.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
Khalim Conn-Kowlessar 2026-06-02 13:36:17 +00:00 • committed by Jun-te Kim
parent a002c7895f
commit ff80fb4b5c
3 changed files with 100 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

15
backend/documents_parser/tests/test_heating_systems_corpus.py
View file

@ -433,9 +433,20 @@ _EXPECTATIONS: tuple[_CorpusExpectation, ...] = (
# 32 code-73 price flip (5.44 → 7.64) per S0380.131's TODO. oil 6
# (B30K) carries a cascade-side residual (HW kWh / SH demand /
# CO2/PE blend) — see open fronts in the post-S0380.168 handover.
#
# Slice S0380.176 closed oil 3 + oil 4 fully via Table 4b combi sub-
# row dispatch in `_table_3a_combi_loss_default_applies`. Pre-slice
# the helper gated only on `main_heating_category` ∈ {1, 2, 3, 6};
# the Elmhurst mapper leaves `main_heating_category=None` on Table
# 4b liquid-fuel boilers, so the cascade fell through to (61)m=0
# despite codes 128/129 being explicit combi sub-rows per SAP 10.2
# Table 4b row names ("Combi oil boiler, pre-/post-1998"). Adding
# the `_TABLE_4B_COMBI_OR_CPSU_CODES` fall-through lands (61)m at
# the spec Table 3a row 1 keep-hot 600 kWh/yr default. Both oil 3
# and oil 4 now EXACT on SAP / cost / CO2 / PE.
_CorpusExpectation(variant='oil 2', block='11a', expected_sap_resid=+0.0000, expected_cost_resid_gbp=+0.0000, expected_co2_resid_kg=+0.0000, expected_pe_resid_kwh=+0.0000),
_CorpusExpectation(variant='oil 3', block='11a', expected_sap_resid=+2.5863, expected_cost_resid_gbp=-61.8906, expected_co2_resid_kg=-14.5815, expected_pe_resid_kwh=-967.0971),
_CorpusExpectation(variant='oil 4', block='11a', expected_sap_resid=+2.5603, expected_cost_resid_gbp=-56.6586, expected_co2_resid_kg=-13.3489, expected_pe_resid_kwh=-884.8990),
_CorpusExpectation(variant='oil 3', block='11a', expected_sap_resid=-0.0000, expected_cost_resid_gbp=-0.0000, expected_co2_resid_kg=+0.0000, expected_pe_resid_kwh=+0.0000),
_CorpusExpectation(variant='oil 4', block='11a', expected_sap_resid=+0.0000, expected_cost_resid_gbp=+0.0000, expected_co2_resid_kg=+0.0000, expected_pe_resid_kwh=+0.0000),
_CorpusExpectation(variant='oil 5', block='11a', expected_sap_resid=+0.0000, expected_cost_resid_gbp=-0.0000, expected_co2_resid_kg=-0.0000, expected_pe_resid_kwh=+0.0000),
_CorpusExpectation(variant='oil 6', block='11a', expected_sap_resid=+3.0518, expected_cost_resid_gbp=-69.7943, expected_co2_resid_kg=-240.6595, expected_pe_resid_kwh=-1112.6558),
_CorpusExpectation(variant='oil pcdb 1', block='11a', expected_sap_resid=+0.0000, expected_cost_resid_gbp=+0.0000, expected_co2_resid_kg=-0.0000, expected_pe_resid_kwh=+0.0000),

16
domain/sap10_calculator/rdsap/cert_to_inputs.py
View file

@ -4794,10 +4794,24 @@ def _table_3a_combi_loss_default_applies(main: Optional[MainHeatingDetail]) -> b
boiler family or a community heat network — outside that set the spec's
"enter '0' if not a combi boiler" rule fires and the cascade must zero
(61)m.
The `main_heating_category` route covers the Open EPC API path where
the cert lodges a SAP 10.2 boiler / heat-network category integer.
The `_TABLE_4B_COMBI_OR_CPSU_CODES` fall-through covers the Elmhurst-
path case where the mapper leaves `main_heating_category=None` but
the cert lodges a Table 4b combi sub-row directly (oil 3 / oil 4 in
heating-systems corpus 001431 — SAP codes 128 / 129 "Combi oil
boiler, pre-/post-1998", FAME fuel — Elmhurst's mapper artifact
leaves the category unset).
"""
if main is None:
return False
return main.main_heating_category in _TABLE_3A_COMBI_LOSS_MAIN_HEATING_CATEGORIES
if main.main_heating_category in _TABLE_3A_COMBI_LOSS_MAIN_HEATING_CATEGORIES:
return True
code = main.sap_main_heating_code
if isinstance(code, int) and code in _TABLE_4B_COMBI_OR_CPSU_CODES:
return True
return False
def _water_heating_worksheet_and_gains(

72
domain/sap10_calculator/rdsap/tests/test_cert_to_inputs.py
View file

@ -4460,6 +4460,78 @@ def test_air_source_heat_pump_main_heating_zeroes_table_3a_combi_loss_per_sap_4_
)
def test_table_4b_combi_oil_boiler_applies_table_3a_combi_loss_per_sap_4_line_7702() -> None:
"""SAP 10.2 §4 line 7702: "Combi loss for each month from Table 3a,
3b or 3c (enter '0' if not a combi boiler)". Table 4b sub-rows 128
("Combi oil boiler, pre-1998") and 129 ("Combi oil boiler, 1998 or
later") are explicitly combi boilers per the Table 4b row names.
Pre-slice `_table_3a_combi_loss_default_applies` gated only on
`main_heating_category` ∈ {1, 2, 3, 6}. The Elmhurst mapper leaves
`main_heating_category=None` on Table 4b liquid-fuel boilers (FAME,
HVO, B30K — codes 128/129/130 combi sub-rows and 124/125/126/127
regular sub-rows). For combi-without-cylinder certs the (61)m
keep-hot default fell through to zero instead of the spec's 600
kWh/yr Table 3a row 1.
Worksheet evidence for heating-systems corpus 001431 oil 3 (Table
4b code 128 + WHC=901 + no cylinder + FAME fuel):
(61)m Jan ≈ 50.96 kWh, annual ≈ 600 kWh/yr
(62)m Jan = 251.39 = 0.85 × (45) + (46) + (61)
(219) HW fuel sum ≈ 3787 kWh/yr (= (62)m / (217)m via Eq D1)
Pre-slice cascade for oil 3: (61)m = 0, (62)m sum = 1935.37 (≡ (45)
sum, off by -600 vs ws 2535.37), (219) = 2876 (off by -911 vs ws
3787).
"""
# Arrange — synthesise oil 3 / oil 4 cert shape: Table 4b code 128
# combi oil boiler (FAME fuel = Table 32 code 73), no cylinder,
# WHC=901 ("HW from main heating"), main_heating_category=None
# (Elmhurst mapper artifact).
combi_main = MainHeatingDetail(
has_fghrs=False,
main_fuel_type=73, # FAME fuel
heat_emitter_type=1,
emitter_temperature=1,
main_heating_control=2106,
main_heating_category=None, # Elmhurst leaves None on Table 4b
sap_main_heating_code=128, # Combi oil boiler, pre-1998
)
epc = make_minimal_sap10_epc(
total_floor_area_m2=_TYPICAL_TFA_M2,
habitable_rooms_count=4,
country_code="ENG",
has_hot_water_cylinder=False, # combi → no cylinder
sap_building_parts=[make_building_part()],
sap_heating=make_sap_heating(
main_heating_details=[combi_main],
water_heating_code=901,
water_heating_fuel=73,
),
)
# Act
wh_result, _ = _water_heating_worksheet_and_gains(
epc=epc,
water_efficiency_pct=0.62, # summer eff for Table 4b code 128
is_instantaneous=False,
primary_age="D",
pcdb_record=None,
)
# Assert — (61)m must be Table 3a row 1 default (600 kWh/yr
# prorated by days-in-month), NOT zero. Sum must land at ~600.
assert wh_result is not None
annual_combi_loss = sum(wh_result.combi_loss_monthly_kwh)
assert abs(annual_combi_loss - 600.0) < 1e-3, (
f"(61)m sum: got {annual_combi_loss!r}, want 600.0 — Table 4b "
f"combi sub-row 128 must apply Table 3a row 1 keep-hot default "
f"per SAP 10.2 §4 line 7702. Pre-slice cascade gated on "
f"main_heating_category ∈ {{1, 2, 3, 6}} only; the Elmhurst-"
f"path Table 4b combi codes (128/129/130) need a fall-through."
)
def test_lighting_co2_factor_blends_table_12a_grid_2_with_table_12d_dual_rate_on_off_peak_certs() -> None:
"""SAP 10.2 Table 12a Grid 2 (PDF p.191) + Table 12d (PDF p.194) —
"other electricity uses" (lighting, pumps + fans, electric shower) on